Adding New Capability in Existing Scientific Application with LLM Assistance

TL;DR

This paper introduces a novel methodology for generating new algorithm code from scratch using large language models, enhancing existing code-translation tools to support innovative algorithm development.

Contribution

It presents a new approach for creating code for unseen algorithms with LLM assistance, extending the capabilities of existing code-translation tools.

Findings

Demonstrates the effectiveness of LLM-assisted code generation for new algorithms.

Enhances the Code-Scribe tool to support novel algorithm implementation.

Shows potential for automating the development of algorithms not present in training data.

Abstract

With the emergence and rapid evolution of large language models (LLM), automating coding tasks has become an important research topic. Many efforts are underway and literature abounds about the efficacy of models and their ability to generate code. A less explored aspect of code generation is for new algorithms, where the training dataset would not have included any previous example of similar code. In this paper we propose a new methodology for writing code from scratch for a new algorithm using LLM assistance, and describe enhancement of a previously developed code-translation tool, Code-Scribe, for new code generation.